Sbp 364 dataset

[vtnphan]

Pull Request

Summary

SBP-XXX Add interaction screening dataset upload endpoint. This introduces a dedicated samplesheet builder and API endpoint for the interaction screening workflow, which requires a structured CSV with sequence file paths and group assignments (query/target → g1/g2).

Changes

• New endpoint POST /datasets/interaction-screening/upload in routes/workflows.py — accepts a list of sequences with id/group and a runId, creates a named Seqera dataset, and uploads the generated samplesheet
• New service function upload_interaction_screening_dataset() in services/datasets.py — builds the samplesheet CSV (id, sequence, group, type columns) and uploads it to Seqera; sequence file paths are resolved under /g/data/yz52/sbp-service/input/interaction_screening/<unique-run-path>/; groups map query → g1, target → g2
• New helper build_unique_dataset_name() in services/datasets.py — generates a timestamped, slug-safe name with a 4-char random suffix; replaces the previous millisecond-timestamp approach
• New schemas SequenceItem (id: str, group: Literal["query", "target"]) and InteractionScreeningDatasetUploadRequest (sequences, runId) in schemas/workflows.py
• Breaking DatasetUploadRequest — removed datasetName and datasetDescription fields; dataset naming is now handled automatically by build_unique_dataset_name
• Breaking create_seqera_dataset() signature changed from (name: str | None, description: str | None) to (name: str = "dataset"); description is no longer sent to the Seqera API
• Security /datasets/upload now requires require_workflow_execution_role (previously unprotected)
• Tests added coverage for the new service function, route, and schemas in test_services_datasets.py, test_additional_coverage.py, and test_schemas.py

How to Test

1. Start the backend locally with valid Seqera env vars (SEQERA_API_URL, SEQERA_ACCESS_TOKEN, WORK_SPACE)

2. Call the new endpoint:

curl -X POST http://localhost:8000/api/workflows/datasets/interaction-screening/upload \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
    "runId": "my-test-run",
    "sequences": [
      {"id": "seq_A", "group": "query"},
      {"id": "seq_B", "group": "target"}
    ]
  }'

3. Verify the response contains a datasetId and "success": true

4. In Seqera Platform, confirm a dataset was created with a uniquely named samplesheet CSV containing the correct g1/g2 groups and file paths under interaction_screening/

5. Run tests:

.venv/bin/python -m pytest tests/test_services_datasets.py tests/test_additional_coverage.py tests/test_schemas.py -v

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation update

Checklist

• I have added tests that prove my fix is effective or that my feature works
• I have added or updated documentation where necessary
• I have run linting and unit tests locally
• The code follows the project's style guidelines

[marius-mather]

Looks good, just a few suggestions for making the code a bit cleaner

[marius-mather]

You shouldn't need to convert this to dicts, just keep it as the pydantic type. That will make the type clearer on upload_interaction_screening_dataset too

[marius-mather]

supply a name explicitly, even if it's just dataset

[marius-mather]

Do all the endpoints under workflows require the workflow execution role? If so, apply it to the router

[marius-mather]

Add a docstring with an example of what the name looks like

[marius-mather]

It's a good habit to call output.close() after you're done with the file (or you can use with io.iStringIO() as output: to automatically close the file after the with: block